1h+ min compiles are not normal. I suspect Goma wasn't working on well on that day, but logs are gone. The reason for 2 hr timeout is primarily for safety, not because CQ team thinks anything below 2hours is good. Our goal is in fact 50% percentile below 1 hour.

Now I see 3 people starring this issue. Do you have recent examples of this?

I might be able to identify the build.
https://goto.google.com/tzohr
http://chromium-build-stats.appspot.com/compiler_proxy_log/2016/04/06/slave299-c4/compiler_proxy.slave299-c4.chrome-bot.log.INFO.20160406-143510.16595.gz
http://chromium-build-stats.appspot.com/ninja_log/2016/04/06/slave299-c4/ninja_log.slave299-c4.chrome-bot.20160406-160018.16579.gz/trace.html

As Andrii mentioned, goma seemed not work well at that time.

Thanks, Yoshisato! Thus, wontfix as it goma outage has been resolved.

Long compile builds are normal for those who change things in base. These compiles can reach 1.5 hours because the world has to be rebuilt.

OK, that's fair.
I don't know what's best way forward. Assigning to decided what to do about this to +Pawel who understands better the reasons behind 2 hour timeout.

If the goal of this issue is making CQ not fail for long compile step, the issue might be already fixed.
If the goal is fixing long compile step, it should be covered by https://bugs.chromium.org/p/chromium/issues/detail?id=667536